Skip to main content
Malaya AI Research Universiti Malaya, FCSIT
Publication

New Research Benchmarks Arabic Fake-News Detection Models

A new PeerJ Computer Science study co-authored by Dr. Aznul and Dr. Norisma compares ensemble and deep-learning approaches to multi-level Arabic fake-news detection.

Staff: Aznul Qalid Md Sabri and Norisma Idris

Using 50,000 Facebook posts from verified Jordanian news agencies, the study evaluated ensemble, recurrent and Transformer models. Extra Trees achieved the strongest overall results, while Transformers showed advantages for fine-grained classification.

View Supporting Source